Over the last 10 years approximately 750 children with medulloblastoma have been treated on consortiabased clinical trials at an estimated cost of over $150 million. Despite this enormous effort, little meaningful molecular data have been generated that will inform the next generation of clinical studies. Consequently, all patients currently receive the same aggressive combination of surgery, radiation and chemotherapy. This treatment inflicts devastating side effects on survivors and fails to cure about one quarter of patients. Using gene expression microarray profiling, we have identified subgroups of human medulloblastoma that display distinct patterns of gene expression, chromosomal alteration, histology and prognosis. The sum of this research suggests that medulloblastoma comprises several subgroups that are likely to require different types or intensities of therapy;however, we still lack the comprehensive understanding of these subgroups necessary to develop new treatments. Work from our group and others has shown that subgroups of brain tumors are generated by cancer stem cells (CSC) that share the gene expression profiles of distinct neural progenitor cells, allowing the identification of their candidate cells-of-origin. Our preliminary data show that two emerging subgroups of medulloblastoma that contain activating mutations in the Sonic hedgehog pathway (from here termed SHH-subgroup) and BETA-CATENIN (CTNNB1-subgroup) display the gene expression profiles of granule neuron precursor cells (GNPC) and precursor cells within the precerebellar neuroepithelium (PCN), respectively. These data suggest the hypothesis that: distinct populations of progenitor cells within the developing hindbrain are predisposed to acquire different gene mutations that transform these into CSC. Since these CSC have distinct cellular origins and molecular properties, then they generate disease subgroups that display different biological and clinical characteristics. We will test this hypothesis by focusing on the SHH and CTTV/VBf-subgroups of medulloblastoma to: (i) develop the first ever spontaneous mouse model of CI/V/VBf-subgroup disease;(ii) Determine if SHH and CTNNB1 -subgroups are generated by distinct types of CSC and associated CSC niches, (iii) Develop approved diagnostic tests of SHH and CTWA/fif-subgroup tumors that can select patients with these tumors for clinical trial, and validate the prognostic significance of CTNNB1- disease in a large prospective clinical trial.